ph1_1sim: Simulation of one Phase I study

In IDDI-BE/PhIdesign: Decision rules and simulation of model-assisted Phase I designs

Simulation of one Phase I study

Description

Simulation of one Phase I study with classical 3+3 design, BOIN, Keyboard or i3+3 design

Usage

ph1_1sim(
  phi,
  phi1 = NULL,
  phi2 = NULL,
  start_dose = 1,
  maxtox = NULL,
  N = NULL,
  truerate = NULL,
  cohortsize = NULL,
  maxN = 9,
  maxNdec = "STAY",
  acc_tit,
  dose_no_titr = NULL,
  BOIN_add33_rule = F,
  design,
  MTD_safer = TRUE,
  seed = NULL,
  halfkey = NULL,
  sim = "NO",
  env = parent.frame(),
  hist = 0
)

Arguments

phi	target toxicity rate (needed for all designs)
phi1	for BOIN/i3+3 design: the highest DLT rate that is deemed subtherapeutic (i.e., underdosing), such that dose escalation should be made, see get_BOIN_rules
phi2	for BOIN/i3+3 design: the lowest DLT rate that is deemed overly toxic (i.e., overdosing), such that dose de-escalation is required, see get_BOIN_rules
start_dose	which dose to start with (could be for instance the second dose level)
maxtox	P(DLT rate>phi|c,n)<=maxtox as value for maxprob in get_Elim_rules
N	number of patients (not needed for design="3+3")
truerate	scenario of true DLT rates for each dose level. This also defines the number of dose levels
cohortsize	cohort size, default is 3
maxN	If N(patients) at next dose level >=maxN, then stop algorithm, default is 9
maxNdec	condition for maxN. Two options: "STAY"= stop at maxN if decision is stay; "ANY"=stop at maxN at any decision
acc_tit	logical indicator if algorithm should start with accelerated titration=algorithm starts at dose level with first DLT
dose_no_titr	first dose at which no accelerated titration anymore
BOIN_add33_rule	logical indicator: modify the decision from de-escalation to stay when observing 1 DLT out of 3 patients
design	"BOIN" or "Keyboard" or "i3+3"
MTD_safer	imposes that the MTD should be for 1)BOIN:<lambda_d, 2)Keyboard:<phi+halfkey, 3)i3+3:<phi2
seed	define seed
halfkey	for Keyboard design: half length of key in keyboard design, see get_Keyboard_rules
sim	if "NO", calculate decision rules, define "YES" if done within simulation program for multiple studies
env	parent environment to pass objects from ph1_sim_OC
hist	complete history of simulation in results

Value

a data.frame with elements

• dose_level: dose level

• dose: actual dose to be used

• incr_ratio: The ratio of successive doses

References

Liu S, Yuan Y. Bayesian Optimal Interval Designs for Phase I Clinical Trials. J R Stat Soc Ser C Appl Stat. 2015;64:507–23 Yan F, Mandrekar SJ, Yuan Y. Keyboard: A Novel Bayesian Toxicity Probability Interval Design for Phase I Clinical Trials. Clin Cancer Res. 2017; 23: 3994–4003

Examples

ph1_1sim(phi=0.3, phi1=0.6*0.3, phi2=1.4*0.3, maxtox=0.95, N=15, 
truerate=c(0.20 ,0.25 ,0.30 ,0.40 ,0.60 ,0.75), cohortsize=3, maxN=9, acc_tit=0, 
design="BOIN", MTD_safer=TRUE)
ph1_1sim(phi=0.3, maxtox=0.95, N=30, 
truerate=c(0.20 ,0.25 ,0.30 ,0.40 ,0.60 ,0.75), cohortsize=3, maxN=9, acc_tit=1, 
design="Keyboard", MTD_safer=TRUE, halfkey=0.05)
ph1_1sim(phi=0.3, phi1=0.6*0.3, phi2=1.4*0.3, maxtox=0.95, N=15, 
truerate=c(0.20 ,0.25 ,0.30 ,0.40 ,0.60 ,0.75), cohortsize=3, maxN=9, acc_tit=0, 
design="i3+3", MTD_safer=TRUE)
ph1_1sim(truerate=c(0.20 ,0.25 ,0.30 ,0.40 ,0.60 ,0.75),acc_tit=1,design="3+3")

IDDI-BE/PhIdesign documentation built on Feb. 7, 2024, 9:32 p.m.